Abstract: This article introduces the Compressed Monte Carlo (CMC) algorithm, and studies one-dimensional and multi-dimensional CMC algorithms. In order to improve the computational efficiency of the particle filter (PF) algorithm, the CMC algorithm and PF algorithm were combined to obtain the Compressed Monte Carlo particle filter (CMC-PF) algorithm with higher computational efficiency. Then, carried out was the application of the CMC-PF algorithm in the structural stiffness parameter identification, of the modified Bouc-Wen (MBW) hysteretic model parameter identification and, of the restoring force model parameter identification of the prefabricated double-column subway elevated station. The research results show that: in the structural stiffness parameter identification, when the particle number is set to 5000 and the rate of compression is set to 0.2, the accuracy of the CMC-PF algorithm is basically the same as that of the PF algorithm, and the calculation time is reduced by 24.13%; in the parameter identification of the MBW model, the accuracy of the CMC-PF algorithm (the particle number is set to 1000, the rate of compression is set to 0.2) is basically the same as that of the PF algorithm (particle number is set to 500), and the calculation time is reduced by 55.52%; and based on the CMC-PF algorithm parameter identification results, restoring force models of the prefabricated double-column subway elevated station are established, which have good accuracy and can reflect the hysteretic performance of the prefabricated double-column subway elevated station. The above research indicates that: in the structural model parameter identification, the CMC-PF algorithm can achieve the accuracy of the PF algorithm and has higher computational efficiency, by selecting both the reasonable particle number and the rate of compression.